nickva commented on issue #2430: Running CouchDB in container and then the 
couchdb pod doesn't release RAM consumed
URL: https://github.com/apache/couchdb/issues/2430#issuecomment-586653684
 
 
   Thanks for stats and the additional info, @fne 
   
   Based on the stats from fabric and `rexi`, it looks like there are interment 
connection errors between nodes in the cluster. It seems some nodes in the 
cluster, get disconnected from other nodes. I was specifically looking at 
`rexi_DOWN` messages, write with quorum errors.
   `rexi` in this case is an internal component which performs inter-node RPC.
   
   These disconnects can also be seen in the original logs as `rexi_buffer : 
cluster unstable` lines. . Those are expected during the initial cluster 
startup and shutdown. 
   
   Monitor to see when those errors start going up and how it relates to RAM 
usage. In other words see if increase in connection errors coincide with 
increased RAM usage.
   
   Also there was bug in 2.3.1 release where sometimes requests like _changes 
feed workers were not cleaned up sometimes after the client connection ends. Is 
your RAM usage increase correlated with specific types of requests?
   
   Another experiment could be, for debugging purposes, to start single node 
cluster and see if it experiences the same issue.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

Reply via email to